A PCE is a computation element in a path computation architecture of a multi-protocol label switching (MPLS) network, which may implement a centralized path computation based on various constraint conditions. A PCC is configured to initiate a path computation request and receive a path computation result, the PCE is configured to receive the path computation request from the PCC, perform a path computation according to the request, and return the path computation result to the PCC. A PCE communication protocol (PCEP) is a protocol for performing the path computation request and reply communications between the PCC and the PCE.
The PCE may be divided into two types: a Stateless PCE and a Stateful PCE. The Stateless PCE only stores network status information which specifically includes network topology information and resource information; while the Stateful PCE will also store information of already-established paths in a network and information of reserved resources used in addition to the network status information. Since the Stateful PCE has more network information, and thus may realize more abundant path computation functions. In an existing path computation method based on the PCEP, after receiving a computation request sent by the PCC, the PCE will compute a new path for the PCC according to the network status information, and when the PCE cannot compute a new path satisfying a condition, the PCE returns computation failure information to the PCC, and the PCC will not acquire an available path.
It can be seen that, the path computation method in the prior art is of poor flexibility, and is not adaptable to the above case, thereby limiting scalability of the network.